Statistical Adjustment for Tactical Choices When Evaluating Team’s Offensive Output Across Five Major European Club Soccer Leagues

Match statistics from the England-France 2022 FIFA World Cup Quarterfinal might suggest England lost despite playing better than France: 16 shot attempts to 8, 5 corners to 2, yet suffering a 1-2 defeat. This interpretation, however, ignores the scoring context.

During the 40 minutes, the match was tied (0-0 and 1-1), France actually led in all of the aforementioned statistical categories. Once ahead, France deliberately ceded the initiative to England for 66 minutes to protect their lead. To study the effects of tactical decisions on offensive outputs like shot attempts and corner kicks, we analyzed sequenced match event data from five major European leagues over 15 years. Our approach incorporates scoring context and other tactical drivers, such as red card differentials and home-field advantage, while controlling for team quality using pre-match betting odds. For that, we leverage modeling approaches tailored towards count response data, priori-tizing balance between quality of fit and simplicity.

Our data analysis provides a thorough confirmation for several intuitive aspects of game dynamics, e.g., that leading or shorthanded teams typically produce less offense, while teams that trail or have more men tend to ramp up their attacks. Beyond this, we develop a statistical adjustment mechanism to teams’ offensive outputs that equalizes the contextual factors for both teams, helping obtain a potentially fairer representation of their relative statistical outputs within a game.

This analysis sheds light on how match context drives observed disparities in offensive outputs and offers an alternative, more nuanced, framework for understanding and assessing team performance.